Accurate prediction of the behavior of internal combustion engines without engine testing can be valuable. First, predicting how an engine will perform under a set of conditions can reveal valuable information without having to incur the cost and time involved in testing an actual engine. Second, expensive prototype refinement is not unusual when testing a new engine design. Predicting behavior of a concept without having to manage prototype development can be a helpful step in engine research accommodated by methods of predicting engine behavior. Accurate engine behavior prediction helps overcome these problems by initially testing new ideas for a new engine design or simulating an existing engine architecture prior to any investment in costly prototype parts or time consuming test cell work. Finally, methods of predicting engine behavior can allow for accelerated testing of a potential system without limitation of manpower and facility availability. While not a substitute for actual testing, accurate prediction can eliminate many otherwise reasonable ideas before an expensive test engine is built or before valuable test cell time is used to test an existing engine.
Additionally, as computer processing power continues to improve, open loop controls using engine prediction data can eliminate both closed loop control methods and reliance on expensive monitoring equipment and the associated costly electrical systems, both of which are prone to wear.
Finally, as emissions controls become increasingly important for environmental reasons, it is valuable to provide a controller for closed loop on-board control adjusting emissions in response to predicted behavior for a set of user demanded operational conditions. Also, as is the case in engine design noted above, a predictive system can help to filter out new designs that fail to meet acceptable emission levels.
Prior art in this area has provided, in general, systems that predict aspects of engine operation but fail to satisfactorily integrate these to provide a predictive engine. By way of example, combustion models have been developed that predict combustion within a combustion chamber on the basis of operator inputted values. The same is true of emissions models, injector models and engine cycle models.
Examples of engine combustion models development are summarized in Heywood, J. B., Internal Combustion Engine Fundamentals (New York: McGraw-Hill, 1988). The combustion model is typically integrated into the engine cycle model. Often accompanying these combustion models are emission formation models that provide an indication of NOx and particulate matter emission. Engine cycle simulation is used to compute thermodynamics and gas dynamics in the engine. GT-Power™ is an example of an engine cycle process modeling module. For engine cycle prediction, several methods have been provided in such systems as StarCD™, Fluent™, FIRE™ and VECTIS™. Hydsim™ software is an example of a prior art injector simulation model. The software can compute injection delay, injection rate shape by modeling details of the injector, including its hydraulic system, mechanical system and electronic system.
Each of these components has been developed, to a greater or lesser extent, but, in general, remain independent models of one aspect of engine behavior. By way of example, consider A. Chow et al., Thermodynamic modeling of complete engine systems—a review (London: Professional Engineering Publishing, 1999) and T. Morel et al., “Virtual Engine/Powertrain/Vehicle” Simulation Tool Solves Complex Interacting System Issues (Warrendale, Pa.: Society of Automotive Engineers International, 2003) 2003-01-0372. In each case, examples of models used to describe aspects of an engine are provided.
Chow discusses the system integration of models which determine engine behavior for a specified rate of combustion. These models quantitatively represent heat transfer, fluid flow, turbo-machinery, and emissions production. As one example of a system integration, Morel describes software which, for a specified rate of combustion, provides a system integration of separate models for heat transfer, fluid flow, turbo-machinery, and hydraulic, electrical, thermal, mechanical, and controls elements of the engine system. Other prediction methods discussed by Morel include components with turbocharger, supercharger and intercooler support; co-simulation options; transient pipe temperature analysis; variable pipe length; multi-user database and dynamic valve motion. The methods taught emphasize the relationship of the engine valves with other components.
Generally, the integrations discussed do not explicitly include the interaction between fuel injection behavior and the rate of combustion. Full engine simulation integration benefits if the method allows for the interdependence of injector performance, combustion pressure and combustion rate.
Generally, limited development in integrated engine predictive methods for direct injection engines has been due, in part, to individual component modeling for direct injection diesel fueled engines that has been satisfactory for the purposes of modeling such engines. For example, the changing conditions inside the combustion chamber caused by combustion in a diesel engine has limited influence on the injection of diesel fuel due, largely, to the extremely high pressures at which diesel fuel is injected, so an integrated model that adjusts the fuel delivery to a cylinder in response to the cylinder pressure does not suffer to any significant extent by ignoring combustion of the fuel in the cylinder. That is, compared to the pressures established in the combustion chamber as a result of combustion events, the pressure at which diesel fuel is injected is many times higher and, therefore, less influenced by changes in combustion chamber pressure.
However, ignoring this part of an engine system for a predictive model can be a problem when a gaseous fuel is used to operate high compression ratio direct injection engines. For diesel engines operating on gaseous fuels, the fuel injection rate depends in general on pressure in the combustion chamber; however the rate of change of cylinder pressure depends on the injection rate, as well as on the turbo-machinery and heat transfer components. Thus for a full system integration separate models for predicting injector behavior and combustion development are important. These models exist independently but have not been adequately integrated into a full system applicable to diesel engines, or engines with diesel engine compression ratios, where those engines operate on gaseous fuels.
Most component models simulate aspects of an engine removed from the operation of the engine system as a whole. The interface between the modeled components and the rest of the system is based on engine test measurements or data. Therefore, the model is not a self-supported full-predictive model. The models are applied to understand the engine process for diagnostic purpose.
As noted above, in injector computation, the model upstream conditions, such as rail pressure and temperature, are from measurements that, in general, do not consider fluctuations in pressure, which might be quite significant in some cases. This is especially important when gas fuel injection is involved. In combustion models, the intake conditions are usually from test measurements on a functioning engine. There is little consideration of the interaction between the injection and the combustion. Likewise, in engine cycle simulation, the combustion process is often derived from the heat release rate measured on an actual engine. This ignores the effect of the intake conditions on the combustion process and simplifies the interaction between injection, combustion and air intake and management processes. It is believed that the interaction between components in the engine system is very important and needs to be well understood in order to shorten engine development phase. However, it is also very complex. Most engine models do not reflect the complex interactions in engine system. Models are detached from the system reducing a model's predictive capability.
Moreover, until recently computer power, has not been adequate to accurately integrate modeling components to predict overall engine behavior in a time frame useful for the purposes contemplated. Developments in computer processing power have provided the necessary computational capacity to model an integrated engine.
Efforts to integrate sub-models together have, in general, not resulted in integrated fuel systems, combustion system and engine breathing systems (as well as emissions systems). In particular, modeling efforts have not been developed for such a predictive integrated system for a gaseous-fueled engine. Also, while the focus of efforts to create engine predictive methods has been directed to the purpose of providing an engine diagnostics tool and predicting engine behavior, little effort has been made on the use of such methods for engine controls.
The present method discloses an integrated engine prediction tool for controlling and predicting engine performance and emissions.